skip to main content

Title: Modeling semantics and pragmatics of spatial prepositions via hierarchical common-sense primitives
Understanding spatial expressions and using them appropriately is necessary for seamless and natural human-machine interaction. However, capturing the semantics and appropriate usage of spatial prepositions is notoriously difficult, because of their vagueness and polysemy. Although modern data-driven approaches are good at capturing statistical regularities in the usage, they usually require substantial sample sizes, often do not generalize well to unseen instances and, most importantly, their structure is essentially opaque to analysis, which makes diagnosing problems and understanding their reasoning process difficult. In this work, we discuss our attempt at modeling spatial senses of prepositions in English using a combination of rule-based and statistical learning approaches. Each preposition model is implemented as a tree where each node computes certain intuitive relations associated with the preposition, with the root computing the final value of the prepositional relation itself. The models operate on a set of artificial 3D “room world” environments, designed in Blender, taking the scene itself as an input. We also discuss our annotation framework used to collect human judgments employed in the model training. Both our factored models and black-box baseline models perform quite well, but the factored models will enable reasoned explanations of spatial relation judgements.
Authors:
; ; ; ; ;
Award ID(s):
1940981
Publication Date:
NSF-PAR ID:
10299975
Journal Name:
Workshop on Spatial Language Understanding and Grounded Communication for Robotics (SpLU-RoboNLP 2021)
Page Range or eLocation-ID:
32-41
Sponsoring Org:
National Science Foundation
More Like this
  1. Spatial patterns in ecology contain useful information about underlying mechanisms and processes. Although there are many summary statistics used to quantify these spatial patterns, there are far fewer models that directly link explicit ecological mechanisms to observed patterns easily derived from available data. We present a model of intraspecific spatial aggregation that quantitatively relates static spatial patterning to negative density dependence. Individuals are placed according to the colonization rule consistent with the Maximum Entropy Theory of Ecology (METE), and die with probability proportional to their abundance raised to a power α, a parameter indicating the degree of density dependence. Thismore »model can therefore be interpreted as a hybridization of MaxEnt and mechanism. Our model shows quantitatively and generally that increasing density dependence randomizes spatial patterning. α = 1 recovers the strongly aggregated METE distribution that is consistent with many ecosystems empirically, and as α → 2 our prediction approaches the binomial distribution consistent with random placement. For 1 < α < 2, our model predicts more aggregation than random placement but less than METE. We additionally relate our mechanistic parameter α to the statistical aggregation parameter k in the negative binomial distribution, giving it an ecological interpretation in the context of density dependence. We use our model to analyze two contrasting datasets, a 50 ha tropical forest and a 64 m 2 serpentine grassland plot. For each dataset, we infer α for individual species as well as a community α parameter. We find that α is generally larger in the tightly packed forest than the sparse grassland, and the degree of density dependence increases at smaller scales. These results are consistent with current understanding in both ecosystems, and we infer this underlying density dependence using only empirical spatial patterns. Our model can easily be applied to other datasets where spatially explicit data are available.« less
  2. We study two approaches for predicting an appropriate pose for a robot to take part in group formations typical of social human conversations subject to the physical layout of the surrounding environment. One method is model-based and explicitly encodes key geometric aspects of conversational formations. The other method is data-driven. It implicitly models key properties of spatial arrangements using graph neural networks and an adversarial training regimen. We evaluate the proposed approaches through quantitative metrics designed for this problem domain and via a human experiment. Our results suggest that the proposed methods are effective at reasoning about the environment layoutmore »and conversational group formations. They can also be used repeatedly to simulate conversational spatial arrangements despite being designed to output a single pose at a time. However, the methods showed different strengths. For example, the geometric approach was more successful at avoiding poses generated in nonfree areas of the environment, but the data-driven method was better at capturing the variability of conversational spatial formations. We discuss ways to address open challenges for the pose generation problem and other interesting avenues for future work.« less
  3. Abstract

    Predicting rain from large-scale environmental variables remains a challenging problem for climate models and it is unclear how well numerical methods can predict the true characteristics of rainfall without smaller (storm) scale information. This study explores the ability of three statistical and machine learning methods to predict 3-hourly rain occurrence and intensity at 0.5° resolution over the tropical Pacific Ocean using rain observations the Global Precipitation Measurement (GPM) satellite radar and large-scale environmental profiles of temperature and moisture from the MERRA-2 reanalysis. We also separated the rain into different types (deep convective, stratiform, and shallow convective) because of theirmore »varying kinematic and thermodynamic structures that might respond to the large-scale environment in different ways. Our expectation was that the popular machine learning methods (i.e., the neural network and random forest) would outperform a standard statistical method (a generalized linear model) because of their more flexible structures, especially in predicting the highly skewed distribution of rain rates for each rain type. However, none of the methods obviously distinguish themselves from one another and each method still has issues with predicting rain too often and not fully capturing the high end of the rain rate distributions, both of which are common problems in climate models. One implication of this study is that machine learning tools must be carefully assessed and are not necessarily applicable to solving all big data problems. Another implication is that traditional climate model approaches are not sufficient to predict extreme rain events and that other avenues need to be pursued.

    « less
  4. Numerical models based on physics represent the state of the art in Earth system modeling and comprise our best tools for generating insights and predictions. Despite rapid growth in computational power, the perceived need for higher model resolutions overwhelms the latest generation computers, reducing the ability of modelers to generate simulations for understanding parameter sensitivities and characterizing variability and uncertainty. Thus, surrogate models are often developed to capture the essential attributes of the full-blown numerical models. Recent successes of machine learning methods, especially deep learning (DL), across many disciplines offer the possibility that complex nonlinear connectionist representations may be ablemore »to capture the underlying complex structures and nonlinear processes in Earth systems. A difficult test for DL-based emulation, which refers to function approximation of numerical models, is to understand whether they can be comparable to traditional forms of surrogate models in terms of computational efficiency while simultaneously reproducing model results in a credible manner. A DL emulation that passes this test may be expected to perform even better than simple models with respect to capturing complex processes and spatiotemporal dependencies. Here, we examine, with a case study in satellite-based remote sensing, the hypothesis that DL approaches can credibly represent the simulations from a surrogate model with comparable computational efficiency. Our results are encouraging in that the DL emulation reproduces the results with acceptable accuracy and often even faster performance. We discuss the broader implications of our results in light of the pace of improvements in high-performance implementations of DL and the growing desire for higher resolution simulations in the Earth sciences.« less
  5. Aslan, Claire (Ed.)
    Abstract The distribution and abundance of plants across the world depends in part on their ability to move, which is commonly characterized by a dispersal kernel. For seeds, the total dispersal kernel (TDK) describes the combined influence of all primary, secondary and higher-order dispersal vectors on the overall dispersal kernel for a plant individual, population, species or community. Understanding the role of each vector within the TDK, and their combined influence on the TDK, is critically important for being able to predict plant responses to a changing biotic or abiotic environment. In addition, fully characterizing the TDK by including allmore »vectors may affect predictions of population spread. Here, we review existing research on the TDK and discuss advances in empirical, conceptual modelling and statistical approaches that will facilitate broader application. The concept is simple, but few examples of well-characterized TDKs exist. We find that significant empirical challenges exist, as many studies do not account for all dispersal vectors (e.g. gravity, higher-order dispersal vectors), inadequately measure or estimate long-distance dispersal resulting from multiple vectors and/or neglect spatial heterogeneity and context dependence. Existing mathematical and conceptual modelling approaches and statistical methods allow fitting individual dispersal kernels and combining them to form a TDK; these will perform best if robust prior information is available. We recommend a modelling cycle to parameterize TDKs, where empirical data inform models, which in turn inform additional data collection. Finally, we recommend that the TDK concept be extended to account for not only where seeds land, but also how that location affects the likelihood of establishing and producing a reproductive adult, i.e. the total effective dispersal kernel.« less